More on Warren Smith's and Anthony Gierzynski's flawed analysis.

by Terry Bouricius // Published March 23, 2009

In my previous blogpost I discussed some of the errors and faulty analysis by Warren Smith, Anthony Gierzynski and Wes Hamilton (though I have the impression that Wes Hamilton "signed-on" rather than performing the faulty analysis himself). In this post I will address additional faulty claims they make.

It should be noted at the outset that their analysis does state that IRV "performed better in this election than plain plurality voting, which (based on top-preference votes) would have elected Wright. That would have been even worse, since Wright actually was a 'lose-to-all loser' among the Big Three, i.e. would have lost head-to-head races versus either Kiss or Montroll." Much of their faulty analysis is therefore focused on showing how IRV is theoretically not as good as other alternative voting methods they apparently prefer to IRV.


Smith and Gierzynski assert that the Burlington election suffered from a "spoiler" scenario. In order to make this claim they need to stretch the usual understanding of the term "spoiler" to fit the data. Normally, the term "spoiler" refers to a candidate who receives relatively few votes, but gets enough to split the majority, throwing the election to a candidate the majority opposes. They argue that the "spoiler" in the Burlington election was the Republican candidate, Kurt Wright, who clearly had an excellent chance of winning, was ahead in the initial tally and then narrowly lost the election in the final runoff to Progressive Bob Kiss. Smith and Gierzynski argue that this Republican front-runner spoiled the chances for the third-place candidate, Democrat, Andy Montroll. I have never seen the term "spoiler" applied to a front-runner "spoiling" the chances of a candidate in third place before. Certainly nobody in Burlington considered this leading candidate to be a "spoiler." That's because people value first choice support, and any candidate strong enough to lead the field in first choice rankings has a credible case to make for running.

This stretch of the meaning of the term "spoiler," so that Smith and Gierzynski can claim there was a spoiler in the Burlington election suggests the term could also be stretched so that the actual winner, Bob Kiss "spoiled" the possible election of the third place finisher Andy Montroll, who would have won, if Kiss hadn't run. The simple fact is that no candidate was considered a "spoiler" by the media or by the voters, and no voters appear to have abandoned their true favorite due to a concern about "spoilers." Under IRV in Burlington, the more similar candidates (based on voter cross ranking), Kiss and Montroll actually avoided the spoiler dynamic that, under plurality election rules (with its spoiler dynamics), would have elected Republican Kurt Wright.

To be fair, Warren and Gierzynzki also use the more appropriate term "favorite betrayal" to describe what could have, but didn't, happen in the Burlington election. They point out that supporters of the Republican candidate could have betrayed their favorite candidate and helped the Democrat win the election if they had voted for the Democrat instead of the Republican, (and many of these Republicans preferred the Democrat to the Progressive). However, such "strategic" calculations played no role in the Burlington election, and there is no reason to think any voter actually abandoned (betrayed) her favorite candidate to help a less preferred choice defeat an even worse choice. Again, as with the false claim of a monotonicity failure (discussed in my previous Blog), Smith and Gierzynski have shown that favorite betrayal could have occurred, but didn't. Because with IRV the viability of any such strategic possibilities are exceedingly difficult to work out in advance, without knowing how all of the other voters are going to rank candidates, voters simply vote sincerely, rather than engaging in such manipulative voting games.

NO "NO-SHOW" PARADOX (or Participation Criterion failure)

Another supposed paradox that Warren Smith and Anthony Gierzynski claim occurred in the Burlington IRV election is the "no-show" or participation paradox. This is the claim that some of Kurt Wright's voters would have had a better outcome if some of them had simply stayed home and not voted at all. To the casual reader this may sound as if they are suggesting that Kurt Wright could have won paradoxically if some of his voters hadn't turned out. But that is not what they are saying. What they are claiming is that if a large number of voters whose favorite candidate was Republican Kurt Wright had not participated, they could have kept Wright from even making it into the final runoff, and allowed the Democrat, Montroll to make it into the final runoff instead. And in that case, Montroll would be elected instead of the Progressive, Bob Kiss.

However, using the actual rankings on the ballots shows that this is not the case in the Burlington election. By removing enough ballots that favored Wright first and also Montroll over Kiss, in order to keep Wright out of the runoff and let Montroll into the final runoff instead, the number of ballots that favored Montroll over Kiss is so diminished that there are not enough ballots left to raise Montroll ahead of Kiss in the final tally. In fact, there wasn't any "no-show" paradox in this election according to the actual ballot data.

But the authors are smart enough to figure out a way to "cook" the data and then assert that this paradox did occur. What they do is assign preferences to voters that the voters themselves did not indicate on their ballots. To be clear, they do not use the actual ballot data, but instead made-up ballot ranking data to manufacture the paradox they apparently so wish had occurred in the Burlington election. Despite the fact that certain voters indicated with their ballots that they had no preference one way or the other between Montroll and Kiss (a plague on both their houses), Gierzynski and Smith feel comfortable "adjusting" the data to show how these voters should have marked their ballots, and then state categorically that there was a no-show paradox. The actual data does not show any such paradox.


As part of Smith-Hamilton-Gierzynski's view that they know what voters would have done, they make a claim that several voting methods in which lower choices count against a voter's own higher choices – Bucklin Voting, Approval Voting and Range Voting – would have elected Montroll in the election. This is unlikely, and flies in the face of the realities of Burlington politics.

Unlike IRV, these methods violate the "later-no-harm criterion" and thus create a clear incentive for campaigns to urge "bullet voting" (ranking only a first choice) and for voters on their own to choose to bullet vote for their top choice candidate – which is why when Bucklin voting was used elsewhere, it was not uncommon for less than 20% of voters to rank a second choice, as opposed to the more than 80% who ranked second choices in Burlington (and higher rates among voters who rightly expected that their first choice was going to be eliminated). In other words, the ballot data would likely look very different due to strategic bullet voting if such a later-harm system had been used. In fact, Bucklin, Approval and Range voting quite possibly would have elected Kurt Wright (the Condorcet-loser among the top three), although this cannot be known for certain -- it is true that Wright would have won if second choice rankings had been randomly reduced in half.

There are other false and misleading statements in the Smith-Hamilton-Gierzynski diatribe, but I will end with a final note about their claims regarding Range Voting. They state that Range Voting does not suffer from "spoilers" and elects the Condorcet-winner in all but rare cases. This is almost certainly false. Specific examples of Range voting spoiler scenarios and other failed criteria are on an election method comparison chart here. Indeed, it is quite possible that a candidate who would lose 55% to 45% in a plurality election might win with range voting, with the chances of other transparently bizarre results all tied to which candidates' backers more effectively gamed the system by insincerely giving no points to anyone but their top choice.

Since Range Voting has never been used for any governmental elections it is hard to be certain how often various paradoxes and pathological outcomes would occur. But it certainly is subject to genuine spoiler scenarios, failure of the later-no-harm criterion, failure to elect Condorcet winners, failure to elect the majority favorite, failure to elect a mutual majority candidate and even the election of the Condorcet-loser (the candidate who would lose every head-to-head contest). See the election method comparison chart for details. One of the most thorough books evaluating various voting methods, Collective Decisions and Voting by Prof. Nicolaus Tideman states (p.238) that Range voting is one of six voting methods that "have defects that are so serious as to disqualify them from consideration."